1. Automatic Performance Tuning Sparse Matrix Algorithms James Demmel www.cs.berkeley.edu/~demmel/cs267_Spr06

2. Berkeley Benchmarking and OPtimization (BeBOP) Prof. Katherine Yelick Rich Vuduc –Some results in this talk are from Vuduc’s PhD thesis, www.cs.berkeley.edu/~richie Rajesh Nishtala, Mark Hoemmen, Hormozd Gahvari, Ankit Jain, Sam Williams, A. Gyulassy S. Kamil, … Eun-Jim Im, Ben Lee, many other earlier contributors bebop.cs.berkeley.edu

3. Outline Motivation for Automatic Performance Tuning Results for sparse matrix kernels OSKI = Optimized Sparse Kernel Interface Tuning Higher Level Algorithms Future Work

4. Motivation for Automatic Performance Tuning Writing high performance software is hard –Make programming easier while getting high speed Ideal: program in your favorite high level language (Matlab, Python, PETSc…) and get a high fraction of peak performance Reality: Best algorithm (and its implementation) can depend strongly on the problem, computer architecture, compiler,… –Best choice can depend on knowing a lot of applied mathematics and computer science How much of this can we teach? How much of this can we automate?

5. Examples of Automatic Performance Tuning (1) Dense BLAS –Sequential –PHiPAC (UCB), then ATLAS (UTK) –Now in Matlab, many other releases –math-atlas.sourceforge.net/ Fast Fourier Transform (FFT) & variations –Sequential and Parallel –FFTW (MIT) –Widely used –www.fftw.org Digital Signal Processing –SPIRAL: www.spiral.net (CMU) Communication Collectives (UCB, UTK) Rose (LLNL), Bernoulli (Cornell), Telescoping Languages (Rice), … More projects, conferences, government reports, …

6. Examples of Automatic Performance Tuning (2) What do dense BLAS, FFTs, signal processing, MPI reductions have in common? –Can do the tuning off-line: once per architecture, algorithm –Can take as much time as necessary (hours, a week…) –At run-time, algorithm choice may depend only on few parameters Matrix dimension, size of FFT, etc. Can’t always do off-line tuning –Algorithm and implementation may strongly depend on data only known at run-time –Ex: Sparse matrix nonzero pattern determines both best data structure and implementation of Sparse-matrix-vector-multiplication (SpMV) –BEBOP project addresses this

7. Tuning Dense BLAS —PHiPAC

8. Tuning Dense BLAS– ATLAS Extends applicability of PHIPAC; Incorporated in Matlab (with rest of LAPACK)

9. Tuning Register Tile Sizes (Dense Matrix Multiply) 333 MHz Sun Ultra 2i 2-D slice of 3-D space; implementations color- coded by performance in Mflop/s 16 registers, but 2-by-3 tile size fastest Needle in a haystack

11. Statistical Models for Automatic Tuning Idea 1: Statistical criterion for stopping a search –A general search model Generate implementation Measure performance Repeat –Stop when probability of being within  of optimal falls below threshold Can estimate distribution on-line Idea 2: Statistical performance models –Problem: Choose 1 among m implementations at run-time –Sample performance off-line, build statistical model

12. Example: Select a Matmul Implementation

13. Example: Support Vector Classification

14. A Sparse Matrix You Encounter Every Day

15. Matrix-vector multiply kernel: y (i)  y (i) + A (i,j) *x (j) for each row i for k=ptr[i] to ptr[i+1] do y[i] = y[i] + val[k]*x[ind[k]] SpMV with Compressed Sparse Row (CSR) Storage Matrix-vector multiply kernel: y (i)  y (i) + A (i,j) *x (j) for each row i for k=ptr[i] to ptr[i+1] do y[i] = y[i] + val[k]*x[ind[k]]

16. Motivation for Automatic Performance Tuning of SpMV Historical trends –Sparse matrix-vector multiply (SpMV): 10% of peak or less Performance depends on machine, kernel, matrix –Matrix known at run-time –Best data structure + implementation can be surprising Our approach: empirical performance modeling and algorithm search

17. SpMV Historical Trends: Fraction of Peak

18. Example: The Difficulty of Tuning n = 21200 nnz = 1.5 M kernel: SpMV Source: NASA structural analysis problem

19. Example: The Difficulty of Tuning n = 21200 nnz = 1.5 M kernel: SpMV Source: NASA structural analysis problem 8x8 dense substructure

20. Taking advantage of block structure in SpMV Bottleneck is time to get matrix from memory –Only 2 flops for each nonzero in matrix Don’t store each nonzero with index, instead store each nonzero r-by-c block with index –Storage drops by up to 2x, if rc >> 1, all 32-bit quantities –Time to fetch matrix from memory decreases Change both data structure and algorithm –Need to pick r and c –Need to change algorithm accordingly In example, is r=c=8 best choice? –Minimizes storage, so looks like a good idea…

21. Speedups on Itanium 2: The Need for Search Reference Best: 4x2 Mflop/s

22. Register Profile: Itanium 2 190 Mflop/s 1190 Mflop/s

23. SpMV Performance (Matrix #2): Generation 2 Ultra 2i - 9%Ultra 3 - 5% Pentium III-M - 15%Pentium III - 19% 63 Mflop/s 35 Mflop/s 109 Mflop/s 53 Mflop/s 96 Mflop/s 42 Mflop/s 120 Mflop/s 58 Mflop/s

24. Register Profiles: Sun and Intel x86 Ultra 2i - 11%Ultra 3 - 5% Pentium III-M - 15%Pentium III - 21% 72 Mflop/s 35 Mflop/s 90 Mflop/s 50 Mflop/s 108 Mflop/s 42 Mflop/s 122 Mflop/s 58 Mflop/s

25. SpMV Performance (Matrix #2): Generation 1 Power3 - 13%Power4 - 14% Itanium 2 - 31%Itanium 1 - 7% 195 Mflop/s 100 Mflop/s 703 Mflop/s 469 Mflop/s 225 Mflop/s 103 Mflop/s 1.1 Gflop/s 276 Mflop/s

26. Register Profiles: IBM and Intel IA-64 Power3 - 17%Power4 - 16% Itanium 2 - 33%Itanium 1 - 8% 252 Mflop/s 122 Mflop/s 820 Mflop/s 459 Mflop/s 247 Mflop/s 107 Mflop/s 1.2 Gflop/s 190 Mflop/s

27. Another example of tuning challenges More complicated non-zero structure in general N = 16614 NNZ = 1.1M

28. Zoom in to top corner More complicated non-zero structure in general N = 16614 NNZ = 1.1M

29. 3x3 blocks look natural, but… More complicated non-zero structure in general Example: 3x3 blocking –Logical grid of 3x3 cells But would lead to lots of “fill-in”

30. Extra Work Can Improve Efficiency! More complicated non-zero structure in general Example: 3x3 blocking –Logical grid of 3x3 cells –Fill-in explicit zeros –Unroll 3x3 block multiplies –“Fill ratio” = 1.5 On Pentium III: 1.5x speedup! –Actual mflop rate 1.5 2 = 2.25 higher

31. Automatic Register Block Size Selection Selecting the r x c block size –Off-line benchmark Precompute Mflops(r,c) using dense A for each r x c Once per machine/architecture –Run-time “search” Sample A to estimate Fill(r,c) for each r x c –Run-time heuristic model Choose r, c to minimize time ~ Fill(r,c) / Mflops(r,c)

32. Accurate and Efficient Adaptive Fill Estimation Idea: Sample matrix –Fraction of matrix to sample: s  [0,1] –Cost ~ O(s * nnz) –Control cost by controlling s Search at run-time: the constant matters! Control s automatically by computing statistical confidence intervals –Idea: Monitor variance Cost of tuning –Lower bound: convert matrix in 5 to 40 unblocked SpMVs –Heuristic: 1 to 11 SpMVs

33. Accuracy of the Tuning Heuristics (1/4) NOTE: “Fair” flops used (ops on explicit zeros not counted as “work”) See p. 375 of Vuduc’s thesis for matrices

34. Accuracy of the Tuning Heuristics (2/4)

35. Accuracy of the Tuning Heuristics (3/4)

36. DGEMV

37. Upper Bounds on Performance for blocked SpMV P = (flops) / (time) –Flops = 2 * nnz(A) Lower bound on time: Two main assumptions –1. Count memory ops only (streaming) –2. Count only compulsory, capacity misses: ignore conflicts Account for line sizes Account for matrix size and nnz Charge minimum access “latency”  i at L i cache &  mem –e.g., Saavedra-Barrera and PMaC MAPS benchmarks

38. Example: L2 Misses on Itanium 2 Misses measured using PAPI [Browne ’00]

39. Example: Bounds on Itanium 2

42. Summary of Other Performance Optimizations Optimizations for SpMV –Register blocking (RB): up to 4x over CSR –Variable block splitting: 2.1x over CSR, 1.8x over RB –Diagonals: 2x over CSR –Reordering to create dense structure + splitting: 2x over CSR –Symmetry: 2.8x over CSR, 2.6x over RB –Cache blocking: 2.8x over CSR –Multiple vectors (SpMM): 7x over CSR –And combinations… Sparse triangular solve –Hybrid sparse/dense data structure: 1.8x over CSR Higher-level kernels –AA T *x, A T A*x: 4x over CSR, 1.8x over RB –A  *x: 2x over CSR, 1.5x over RB

43. Example: Sparse Triangular Factor Raefsky4 (structural problem) + SuperLU + colmmd N=19779, nnz=12.6 M Dense trailing triangle: dim=2268, 20% of total nz Can be as high as 90+%! 1.8x over CSR

44. Cache Optimizations for AA T *x Cache-level: Interleave multiplication by A, A T –Only fetch A from memory once Register-level: a i T to be r  c block row, or diag row dot product “axpy” … …

45. Example: Combining Optimizations Register blocking, symmetry, multiple (k) vectors –Three low-level tuning parameters: r, c, v v k X Y A c r += *

46. Example: Combining Optimizations Register blocking, symmetry, and multiple vectors [Ben Lee @ UCB] –Symmetric, blocked, 1 vector Up to 2.6x over nonsymmetric, blocked, 1 vector –Symmetric, blocked, k vectors Up to 2.1x over nonsymmetric, blocked, k vecs. Up to 7.3x over nonsymmetric, nonblocked, 1, vector –Symmetric Storage: up to 64.7% savings

47. Potential Impact on Applications: T3P Application: accelerator design [Ko] 80% of time spent in SpMV Relevant optimization techniques –Symmetric storage –Register blocking On Single Processor Itanium 2 –1.68x speedup 532 Mflops, or 15% of 3.6 GFlop peak –4.4x speedup with multiple (8) vectors 1380 Mflops, or 38% of peak

48. Potential Impact on Applications: Omega3P Application: accelerator cavity design [Ko] Relevant optimization techniques –Symmetric storage –Register blocking –Reordering Reverse Cuthill-McKee ordering to reduce bandwidth Traveling Salesman Problem-based ordering to create blocks –Nodes = columns of A –Weights(u, v) = no. of nz u, v have in common –Tour = ordering of columns –Choose maximum weight tour –See [Pinar & Heath ’97] 2.1x speedup on Power 4, but SPMV not dominant

49. Source: Accelerator Cavity Design Problem (Ko via Husbands)

50. 100x100 Submatrix Along Diagonal

51. Post-RCM Reordering

52. Before: Green + Red After: Green + Blue “Microscopic” Effect of RCM Reordering

53. “Microscopic” Effect of Combined RCM+TSP Reordering Before: Green + Red After: Green + Blue

54. (Omega3P)

55. Optimized Sparse Kernel Interface - OSKI Provides sparse kernels automatically tuned for user’s matrix & machine –BLAS-style functionality: SpMV, Ax & A T y, TrSV –Hides complexity of run-time tuning –Includes new, faster locality-aware kernels: A T Ax, A k x Faster than standard implementations –Up to 4x faster matvec, 1.8x trisolve, 4x A T A*x For “advanced” users & solver library writers –Available as stand-alone library (OSKI 1.0.1b, 3/06) –Available as PETSc extension (OSKI-PETSc.1d, 3/06) –Bebop.cs.berkeley.edu/oski

56. How the OSKI Tunes (Overview) Benchmark data 1. Build for Target Arch. 2. Benchmark Heuristic models 1. Evaluate Models Generated code variants 2. Select Data Struct. & Code Library Install-Time (offline) Application Run-Time To user: Matrix handle for kernel calls Workload from program monitoring Extensibility: Advanced users may write & dynamically add “Code variants” and “Heuristic models” to system. History Matrix

57. How the OSKI Tunes (Overview) At library build/install-time –Pre-generate and compile code variants into dynamic libraries –Collect benchmark data Measures and records speed of possible sparse data structure and code variants on target architecture –Installation process uses standard, portable GNU AutoTools At run-time –Library “tunes” using heuristic models Models analyze user’s matrix & benchmark data to choose optimized data structure and code –Non-trivial tuning cost: up to ~40 mat-vecs Library limits the time it spends tuning based on estimated workload –provided by user or inferred by library User may reduce cost by saving tuning results for application on future runs with same or similar matrix

58. Optimizations in the Initial OSKI Release Fully automatic heuristics for –Sparse matrix-vector multiply Register-level blocking Register-level blocking + symmetry + multiple vectors Cache-level blocking –Sparse triangular solve with register-level blocking and “switch-to-dense” optimization –Sparse A T A*x with register-level blocking User may select other optimizations manually –Diagonal storage optimizations, reordering, splitting; tiled matrix powers kernel ( A k *x ) –All available in dynamic libraries –Accessible via high-level embedded script language “Plug-in” extensibility –Very advanced users may write their own heuristics, create new data structures/code variants and dynamically add them to the system

59. How to Call OSKI: Basic Usage May gradually migrate existing apps –Step 1: “Wrap” existing data structures –Step 2: Make BLAS-like kernel calls int* ptr = …, *ind = …; double* val = …; /* Matrix, in CSR format */ double* x = …, *y = …; /* Let x and y be two dense vectors */ /* Compute y =  ·y +  ·A·x, 500 times */ for( i = 0; i < 500; i++ ) my_matmult( ptr, ind, val, , x, , y );

60. How to Call OSKI: Basic Usage May gradually migrate existing apps –Step 1: “Wrap” existing data structures –Step 2: Make BLAS-like kernel calls int* ptr = …, *ind = …; double* val = …; /* Matrix, in CSR format */ double* x = …, *y = …; /* Let x and y be two dense vectors */ /* Step 1: Create OSKI wrappers around this data */ oski_matrix_t A_tunable = oski_CreateMatCSR(ptr, ind, val, num_rows, num_cols, SHARE_INPUTMAT, …); oski_vecview_t x_view = oski_CreateVecView(x, num_cols, UNIT_STRIDE); oski_vecview_t y_view = oski_CreateVecView(y, num_rows, UNIT_STRIDE); /* Compute y =  ·y +  ·A·x, 500 times */ for( i = 0; i < 500; i++ ) my_matmult( ptr, ind, val, , x, , y );

61. How to Call OSKI: Basic Usage May gradually migrate existing apps –Step 1: “Wrap” existing data structures –Step 2: Make BLAS-like kernel calls int* ptr = …, *ind = …; double* val = …; /* Matrix, in CSR format */ double* x = …, *y = …; /* Let x and y be two dense vectors */ /* Step 1: Create OSKI wrappers around this data */ oski_matrix_t A_tunable = oski_CreateMatCSR(ptr, ind, val, num_rows, num_cols, SHARE_INPUTMAT, …); oski_vecview_t x_view = oski_CreateVecView(x, num_cols, UNIT_STRIDE); oski_vecview_t y_view = oski_CreateVecView(y, num_rows, UNIT_STRIDE); /* Compute y =  ·y +  ·A·x, 500 times */ for( i = 0; i < 500; i++ ) oski_MatMult(A_tunable, OP_NORMAL, , x_view, , y_view); /* Step 2 */

62. How to Call OSKI: Tune with Explicit Hints User calls “tune” routine –May provide explicit tuning hints (OPTIONAL) oski_matrix_t A_tunable = oski_CreateMatCSR( … ); /* … */ /* Tell OSKI we will call SpMV 500 times (workload hint) */ oski_SetHintMatMult(A_tunable, OP_NORMAL, , x_view, , y_view, 500); /* Tell OSKI we think the matrix has 8x8 blocks (structural hint) */ oski_SetHint(A_tunable, HINT_SINGLE_BLOCKSIZE, 8, 8); oski_TuneMat(A_tunable); /* Ask OSKI to tune */ for( i = 0; i < 500; i++ ) oski_MatMult(A_tunable, OP_NORMAL, , x_view, , y_view);

63. How the User Calls OSKI: Implicit Tuning Ask library to infer workload –Library profiles all kernel calls –May periodically re-tune oski_matrix_t A_tunable = oski_CreateMatCSR( … ); /* … */ for( i = 0; i < 500; i++ ) { oski_MatMult(A_tunable, OP_NORMAL, , x_view, , y_view); oski_TuneMat(A_tunable); /* Ask OSKI to tune */ }

64. Quick-and-dirty Parallelism: OSKI-PETSc Extend PETSc’s distributed memory SpMV (MATMPIAIJ) p0 p1 p2 p3 PETSc –Each process stores diag (all-local) and off-diag submatrices OSKI-PETSc: –Add OSKI wrappers –Each submatrix tuned independently

65. OSKI-PETSc Proof-of-Concept Results Matrix 1: Accelerator cavity design (R. Lee @ SLAC) –N ~ 1 M, ~40 M non-zeros –2x2 dense block substructure –Symmetric Matrix 2: Linear programming (Italian Railways) –Short-and-fat: 4k x 1M, ~11M non-zeros –Highly unstructured –Big speedup from cache-blocking: no native PETSc format Evaluation machine: Xeon cluster –Peak: 4.8 Gflop/s per node

66. Accelerator Cavity Matrix

67. OSKI-PETSc Performance: Accel. Cavity

68. Linear Programming Matrix …

69. OSKI-PETSc Performance: LP Matrix

70. Tuning Higher Level Algorithms So far we have tuned a single sparse matrix kernel y = A T *A*x motivated by higher level algorithm (SVD) What can we do by extending tuning to a higher level? Consider Krylov subspace methods for Ax=b, Ax = x Conjugate Gradients (CG), GMRES, Lanczos, … Inner loop does y=A*x, dot products, saxpys, scalar ops Inner loop costs at least O(1) messages k iterations cost at least O(k) messages Our goal: show how to do k iterations with O(1) messages Possible payoff – make Krylov subspace methods much faster on machines with slow networks Memory bandwidth improvements too (not discussed) Obstacles: numerical stability, preconditioning, …

71. Krylov Subspace Methods for Solving Ax=b Compute a basis for a subspace V by doing y = A*x k times Find “best” solution in that Krylov subspace V Given starting vector x 1, V spanned by x 2 = A*x 1, x 3 = A*x 2, …, x k = A*x k-1 GMRES finds an orthogonal basis of V by Gram-Schmidt, so it actually does a different set of SpMVs than in last bullet

72. Example: Standard GMRES r = b - A*x 1,  = length(r), v 1 = r /  … length(r) = sqrt(  r i 2 ) for m= 1 to k do w = A*v m … at least O(1) messages for i = 1 to m do … Gram-Schmidt h im = dotproduct(v i, w ) … at least O(1) messages, or log(p) w = w – h im * v i end for h m+1,m = length(w) … at least O(1) messages, or log(p) v m+1 = w / h m+1,m end for find y minimizing length( H k * y –  e 1 ) … small, local problem new x = x 1 + V k * y … V k = [v 1, v 2, …, v k ] O(k 2 ), or O(k 2 log p ), messages altogether

73. Example: Computing [Ax,A 2 x,A 3 x,…,A k x] for A tridiagonal Different basis for same Krylov subspace What can Proc 1 compute without communication? x(1:30): (Ax)(1:30): (A 2 x)(1:30): (A 8 x)(1:30): Proc 1 Proc 2...

74. Example: Computing [Ax,A 2 x,A 3 x,…,A k x] for A tridiagonal x(1:30): (Ax)(1:30): (A 2 x)(1:30):... (A 8 x)(1:30): Proc 1 Proc 2 Computing missing entries with 1 communication, redundant work

75. x(1:30): (Ax)(1:30): (A 2 x)(1:30):... (A 8 x)(1:30): Proc 1 Proc 2 Saving half the redundant work Example: Computing [Ax,A 2 x,A 3 x,…,A k x] for A tridiagonal

76. Example: Computing [Ax,A 2 x,A 3 x,…,A k x] for Laplacian A = 5pt Laplacian in 2D, Communicated point for k=3 shown

77. Latency-Avoiding GMRES (1) r = b - A*x 1,  = length(r), v 1 = r /  …  O(log p) messages W k+1 = [v 1, A * v 1, A 2 * v 1, …, A k * v 1 ] … O(1) messages [Q, R] = qr(W k+1 ) … QR decomposition, O(log p) messages H k = R(:, 2:k+1) * (R(1:k,1:k)) -1 … small, local problem find y minimizing length( H k * y –  e 1 ) … small, local problem new x = x 1 + Q k * y … local problem O(log p) messages altogether Independent of k

78. Latency-Avoiding GMRES (2) [Q, R] = qr(W k+1 ) … QR decomposition, O(log p) messages Easy, but not so stable way to do it: X(myproc) = W k+1 T (myproc) * W k+1 (myproc) … local computation Y = sum_reduction(X(myproc)) … O(log p) messages … Y = W k+1 T * W k+1 R = (cholesky(Y)) T … small, local computation Q(myproc) = W k+1 (myproc) * R -1 … local computation

79. Numerical example (1) Diagonal matrix with n=1000, A ii from 1 down to 10 -5 Instability as k grows, after many iterations

80. Numerical Example (2) Partial remedy: restarting periodically (every 120 iterations) Other remedies: high precision, different basis than [v, A * v, …, A k * v ]

81. Operation Counts for [Ax,A 2 x,A 3 x,…,A k x] on p procs ProblemPer-proc costStandardOptimized 1D mesh#messages2k2 (tridiagonal)#words sent2k #flops5kn/p 5kn/p + 5k 2 memory(k+4)n/p(k+4)n/p + 8k 3D mesh#messages26k26 27 pt stencil#words sent 6kn 2 p -2/3 + 12knp -1/3 + O(k) 6kn 2 p -2/3 + 12k 2 np -1/3 + O(k 3 ) #flops 53kn 3 /p53kn 3 /p + O(k 2 n 2 p -2/3 ) memory (k+28)n 3 /p+ 6n 2 p -2/3 + O(np -1/3 ) (k+28)n 3 /p + 168kn 2 p -2/3 + O(k 2 np -1/3 )

82. Summary and Future Work Dense –LAPACK –ScaLAPACK Communication primitives Sparse –Kernels, Stencils –Higher level algorithms All of the above on new architectures –Vector, SMPs, multicore, Cell, … High level support for tuning –Specification language –Integration into compilers

83. Extra Slides

84. A Sparse Matrix You Encounter Every Day Who am I? I am a Big Repository Of useful And useless Facts alike. Who am I? (Hint: Not your e-mail inbox.)

85. What about the Google Matrix? Google approach –Approx. once a month: rank all pages using connectivity structure Find dominant eigenvector of a matrix –At query-time: return list of pages ordered by rank Matrix: A =  G + (1-  )(1/n)uu T –Markov model: Surfer follows link with probability , jumps to a random page with probability 1-  –G is n x n connectivity matrix [n  billions] g ij is non-zero if page i links to page j Normalized so each column sums to 1 Very sparse: about 7—8 non-zeros per row (power law dist.) –u is a vector of all 1 values –Steady-state probability x i of landing on page i is solution to x = Ax Approximate x by power method: x = A k x 0 –In practice, k  25

86. Current Work Public software release Impact on library designs: Sparse BLAS, Trilinos, PETSc, … Integration in large-scale applications –DOE: Accelerator design; plasma physics –Geophysical simulation based on Block Lanczos ( A T A*X ; LBL) Systematic heuristics for data structure selection? Evaluation of emerging architectures –Revisiting vector micros Other sparse kernels –Matrix triple products, A k *x Parallelism Sparse benchmarks (with UTK) [Gahvari & Hoemmen] Automatic tuning of MPI collective ops [Nishtala, et al.]

87. Summary of High-Level Themes “Kernel-centric” optimization –Vs. basic block, trace, path optimization, for instance –Aggressive use of domain-specific knowledge Performance bounds modeling –Evaluating software quality –Architectural characterizations and consequences Empirical search –Hybrid on-line/run-time models Statistical performance models –Exploit information from sampling, measuring

88. Related Work My bibliography: 337 entries so far Sample area 1: Code generation –Generative & generic programming –Sparse compilers –Domain-specific generators Sample area 2: Empirical search-based tuning –Kernel-centric linear algebra, signal processing, sorting, MPI, … –Compiler-centric profiling + FDO, iterative compilation, superoptimizers, self- tuning compilers, continuous program optimization

89. Future Directions (A Bag of Flaky Ideas) Composable code generators and search spaces New application domains –PageRank: multilevel block algorithms for topic-sensitive search? New kernels: cryptokernels –rich mathematical structure germane to performance; lots of hardware New tuning environments –Parallel, Grid, “whole systems” Statistical models of application performance –Statistical learning of concise parametric models from traces for architectural evaluation –Compiler/automatic derivation of parametric models

90. Possible Future Work Different Architectures –New FP instruction sets (SSE2) –SMP / multicore platforms –Vector architectures Different Kernels Higher Level Algorithms –Parallel versions of kenels, with optimized communication –Block algorithms (eg Lanczos) –XBLAS (extra precision) Produce Benchmarks –Augment HPCC Benchmark Make it possible to combine optimizations easily for any kernel Related tuning activities (LAPACK & ScaLAPACK)

91. Review of Tuning by Illustration (Extra Slides)

92. Splitting for Variable Blocks and Diagonals Decompose A = A 1 + A 2 + … A t –Detect “canonical” structures (sampling) –Split –Tune each A i –Improve performance and save storage New data structures –Unaligned block CSR Relax alignment in rows & columns –Row-segmented diagonals

93. Example: Variable Block Row (Matrix #12) 2.1x over CSR 1.8x over RB

94. Example: Row-Segmented Diagonals 2x over CSR

95. Mixed Diagonal and Block Structure

96. Summary Automated block size selection –Empirical modeling and search –Register blocking for SpMV, triangular solve, A T A*x Not fully automated –Given a matrix, select splittings and transformations Lots of combinatorial problems –TSP reordering to create dense blocks (Pinar ’97; Moon, et al. ’04)

97. Extra Slides

98. A Sparse Matrix You Encounter Every Day Who am I? I am a Big Repository Of useful And useless Facts alike. Who am I? (Hint: Not your e-mail inbox.)

99. Problem Context Sparse kernels abound –Models of buildings, cars, bridges, economies, … –Google PageRank algorithm Historical trends –Sparse matrix-vector multiply (SpMV): 10% of peak –2x faster with “hand-tuning” –Tuning becoming more difficult over time –Promise of automatic tuning: PHiPAC/ATLAS, FFTW, … Challenges to high-performance –Not dense linear algebra! Complex data structures: indirect, irregular memory access Performance depends strongly on run-time inputs

100. Key Questions, Ideas, Conclusions How to tune basic sparse kernels automatically? –Empirical modeling and search Up to 4x speedups for SpMV 1.8x for triangular solve 4x for A T A*x; 2x for A 2 *x 7x for multiple vectors What are the fundamental limits on performance? –Kernel-, machine-, and matrix-specific upper bounds Achieve 75% or more for SpMV, limiting low-level tuning Consequences for architecture? General techniques for empirical search-based tuning? –Statistical models of performance

101. Road Map Sparse matrix-vector multiply (SpMV) in a nutshell Historical trends and the need for search Automatic tuning techniques Upper bounds on performance Statistical models of performance

102. Matrix-vector multiply kernel: y (i)  y (i) + A (i,j) *x (j) for each row i for k=ptr[i] to ptr[i+1] do y[i] = y[i] + val[k]*x[ind[k]] Compressed Sparse Row (CSR) Storage Matrix-vector multiply kernel: y (i)  y (i) + A (i,j) *x (j) for each row i for k=ptr[i] to ptr[i+1] do y[i] = y[i] + val[k]*x[ind[k]]

103. Road Map Sparse matrix-vector multiply (SpMV) in a nutshell Historical trends and the need for search Automatic tuning techniques Upper bounds on performance Statistical models of performance

104. Historical Trends in SpMV Performance The Data –Uniprocessor SpMV performance since 1987 –“Untuned” and “Tuned” implementations –Cache-based superscalar micros; some vectors –LINPACK

105. SpMV Historical Trends: Mflop/s

106. Example: The Difficulty of Tuning n = 21216 nnz = 1.5 M kernel: SpMV Source: NASA structural analysis problem

107. Still More Surprises More complicated non-zero structure in general

108. Still More Surprises More complicated non-zero structure in general Example: 3x3 blocking –Logical grid of 3x3 cells

109. Historical Trends: Mixed News Observations –Good news: Moore’s law like behavior –Bad news: “Untuned” is 10% peak or less, worsening –Good news: “Tuned” roughly 2x better today, and improving –Bad news: Tuning is complex –(Not really news: SpMV is not LINPACK) Questions –Application: Automatic tuning? –Architect: What machines are good for SpMV?

110. Road Map Sparse matrix-vector multiply (SpMV) in a nutshell Historical trends and the need for search Automatic tuning techniques –SpMV [SC’02; IJHPCA ’04b] –Sparse triangular solve (SpTS) [ICS/POHLL ’02] –A T A*x [ICCS/WoPLA ’03] Upper bounds on performance Statistical models of performance

111. SPARSITY: Framework for Tuning SpMV SPARSITY: Automatic tuning for SpMV [Im & Yelick ’99] –General approach Identify and generate implementation space Search space using empirical models & experiments –Prototype library and heuristic for choosing register block size Also: cache-level blocking, multiple vectors What’s new? –New block size selection heuristic Within 10% of optimal — replaces previous version –Expanded implementation space Variable block splitting, diagonals, combinations –New kernels: sparse triangular solve, A T A*x, A  *x

112. Automatic Register Block Size Selection Selecting the r x c block size –Off-line benchmark: characterize the machine Precompute Mflops(r,c) using dense matrix for each r x c Once per machine/architecture –Run-time “search”: characterize the matrix Sample A to estimate Fill(r,c) for each r x c –Run-time heuristic model Choose r, c to maximize Mflops(r,c) / Fill(r,c) Run-time costs –Up to ~40 SpMVs (empirical worst case)

113. Accuracy of the Tuning Heuristics (1/4) NOTE: “Fair” flops used (ops on explicit zeros not counted as “work”) DGEMV

114. Accuracy of the Tuning Heuristics (2/4) DGEMV

115. Accuracy of the Tuning Heuristics (3/4) DGEMV

116. Accuracy of the Tuning Heuristics (4/4) DGEMV

117. Road Map Sparse matrix-vector multiply (SpMV) in a nutshell Historical trends and the need for search Automatic tuning techniques Upper bounds on performance –SC’02 Statistical models of performance

118. Motivation for Upper Bounds Model Questions –Speedups are good, but what is the speed limit? Independent of instruction scheduling, selection –What machines are “good” for SpMV?

119. Upper Bounds on Performance: Blocked SpMV P = (flops) / (time) –Flops = 2 * nnz(A) Lower bound on time: Two main assumptions –1. Count memory ops only (streaming) –2. Count only compulsory, capacity misses: ignore conflicts Account for line sizes Account for matrix size and nnz Charge min access “latency”  i at L i cache &  mem –e.g., Saavedra-Barrera and PMaC MAPS benchmarks

120. Example: Bounds on Itanium 2

123. Fraction of Upper Bound Across Platforms

124. Achieved Performance and Machine Balance Machine balance [Callahan ’88; McCalpin ’95] –Balance = Peak Flop Rate / Bandwidth (flops / double) Ideal balance for mat-vec:  2 flops / double –For SpMV, even less SpMV ~ streaming –1 / (avg load time to stream 1 array) ~ (bandwidth) –“Sustained” balance = peak flops / model bandwidth

126. Where Does the Time Go? Most time assigned to memory Caches “disappear” when line sizes are equal –Strictly increasing line sizes

127. Execution Time Breakdown: Matrix 40

128. Speedups with Increasing Line Size

129. Summary: Performance Upper Bounds What is the best we can do for SpMV? –Limits to low-level tuning of blocked implementations –Refinements? What machines are good for SpMV? –Partial answer: balance characterization Architectural consequences? –Example: Strictly increasing line sizes

130. Road Map Sparse matrix-vector multiply (SpMV) in a nutshell Historical trends and the need for search Automatic tuning techniques Upper bounds on performance Tuning other sparse kernels Statistical models of performance –FDO ’00; IJHPCA ’04a

131. Statistical Models for Automatic Tuning Idea 1: Statistical criterion for stopping a search –A general search model Generate implementation Measure performance Repeat –Stop when probability of being within  of optimal falls below threshold Can estimate distribution on-line Idea 2: Statistical performance models –Problem: Choose 1 among m implementations at run-time –Sample performance off-line, build statistical model

132. Example: Select a Matmul Implementation

133. Example: Support Vector Classification

134. Road Map Sparse matrix-vector multiply (SpMV) in a nutshell Historical trends and the need for search Automatic tuning techniques Upper bounds on performance Tuning other sparse kernels Statistical models of performance Summary and Future Work

135. Summary of High-Level Themes “Kernel-centric” optimization –Vs. basic block, trace, path optimization, for instance –Aggressive use of domain-specific knowledge Performance bounds modeling –Evaluating software quality –Architectural characterizations and consequences Empirical search –Hybrid on-line/run-time models Statistical performance models –Exploit information from sampling, measuring

136. Related Work My bibliography: 337 entries so far Sample area 1: Code generation –Generative & generic programming –Sparse compilers –Domain-specific generators Sample area 2: Empirical search-based tuning –Kernel-centric linear algebra, signal processing, sorting, MPI, … –Compiler-centric profiling + FDO, iterative compilation, superoptimizers, self- tuning compilers, continuous program optimization

137. Future Directions (A Bag of Flaky Ideas) Composable code generators and search spaces New application domains –PageRank: multilevel block algorithms for topic-sensitive search? New kernels: cryptokernels –rich mathematical structure germane to performance; lots of hardware New tuning environments –Parallel, Grid, “whole systems” Statistical models of application performance –Statistical learning of concise parametric models from traces for architectural evaluation –Compiler/automatic derivation of parametric models

138. Acknowledgements Super-advisors: Jim and Kathy Undergraduate R.A.s: Attila, Ben, Jen, Jin, Michael, Rajesh, Shoaib, Sriram, Tuyet-Linh See pages xvi—xvii of dissertation.

139. TSP-based Reordering: Before (Pinar ’97; Moon, et al ‘04)

140. TSP-based Reordering: After (Pinar ’97; Moon, et al ‘04) Up to 2x speedups over CSR

141. Example: L2 Misses on Itanium 2 Misses measured using PAPI [Browne ’00]

142. Example: Distribution of Blocked Non-Zeros

143. Register Profile: Itanium 2 190 Mflop/s 1190 Mflop/s

144. Register Profiles: Sun and Intel x86 Ultra 2i - 11%Ultra 3 - 5% Pentium III-M - 15%Pentium III - 21% 72 Mflop/s 35 Mflop/s 90 Mflop/s 50 Mflop/s 108 Mflop/s 42 Mflop/s 122 Mflop/s 58 Mflop/s

145. Register Profiles: IBM and Intel IA-64 Power3 - 17%Power4 - 16% Itanium 2 - 33%Itanium 1 - 8% 252 Mflop/s 122 Mflop/s 820 Mflop/s 459 Mflop/s 247 Mflop/s 107 Mflop/s 1.2 Gflop/s 190 Mflop/s

146. Accurate and Efficient Adaptive Fill Estimation Idea: Sample matrix –Fraction of matrix to sample: s  [0,1] –Cost ~ O(s * nnz) –Control cost by controlling s Search at run-time: the constant matters! Control s automatically by computing statistical confidence intervals –Idea: Monitor variance Cost of tuning –Lower bound: convert matrix in 5 to 40 unblocked SpMVs –Heuristic: 1 to 11 SpMVs

147. Sparse/Dense Partitioning for SpTS Partition L into sparse (L 1,L 2 ) and dense L D : Perform SpTS in three steps: Sparsity optimizations for (1)—(2); DTRSV for (3) Tuning parameters: block size, size of dense triangle

148. SpTS Performance: Power3

150. Summary of SpTS and AA T *x Results SpTS — Similar to SpMV –1.8x speedups; limited benefit from low-level tuning AA T x, A T Ax –Cache interleaving only: up to 1.6x speedups –Reg + cache: up to 4x speedups 1.8x speedup over register only –Similar heuristic; same accuracy (~ 10% optimal) –Further from upper bounds: 60—80% Opportunity for better low-level tuning a la PHiPAC/ATLAS Matrix triple products? A k *x ? –Preliminary work

151. Register Blocking: Speedup

152. Register Blocking: Performance

153. Register Blocking: Fraction of Peak

154. Example: Confidence Interval Estimation

155. Costs of Tuning

156. Splitting + UBCSR: Pentium III

157. Splitting + UBCSR: Power4

158. Splitting+UBCSR Storage: Power4

160. Example: Variable Block Row (Matrix #13)

161. Dense Tuning is Hard, Too Even dense matrix multiply can be notoriously difficult to tune

162. Dense matrix multiply: surprising performance as register tile size varies.

164. Preliminary Results (Matrix Set 2): Itanium 2 Web/IR DenseFEMFEM (var)BioLPEconStat

165. Multiple Vector Performance

167. What about the Google Matrix? Google approach –Approx. once a month: rank all pages using connectivity structure Find dominant eigenvector of a matrix –At query-time: return list of pages ordered by rank Matrix: A =  G + (1-  )(1/n)uu T –Markov model: Surfer follows link with probability , jumps to a random page with probability 1-  –G is n x n connectivity matrix [n  3 billion] g ij is non-zero if page i links to page j Normalized so each column sums to 1 Very sparse: about 7—8 non-zeros per row (power law dist.) –u is a vector of all 1 values –Steady-state probability x i of landing on page i is solution to x = Ax Approximate x by power method: x = A k x 0 –In practice, k  25

169. MAPS Benchmark Example: Power4

170. MAPS Benchmark Example: Itanium 2

171. Saavedra-Barrera Example: Ultra 2i

173. Summary of Results: Pentium III

174. Summary of Results: Pentium III (3/3)

175. Execution Time Breakdown (PAPI): Matrix 40

176. Preliminary Results (Matrix Set 1): Itanium 2 LPFEMFEM (var)AssortedDense

177. Tuning Sparse Triangular Solve (SpTS) Compute x=L -1 *b where L sparse lower triangular, x & b dense L from sparse LU has rich dense substructure –Dense trailing triangle can account for 20—90% of matrix non-zeros SpTS optimizations –Split into sparse trapezoid and dense trailing triangle –Use tuned dense BLAS (DTRSV) on dense triangle –Use Sparsity register blocking on sparse part Tuning parameters –Size of dense trailing triangle –Register block size

178. Sparse Kernels and Optimizations Kernels –Sparse matrix-vector multiply (SpMV): y=A*x –Sparse triangular solve (SpTS): x=T -1 *b –y=AA T *x, y=A T A*x –Powers (y=A k *x), sparse triple-product (R*A*R T ), … Optimization techniques (implementation space) –Register blocking –Cache blocking –Multiple dense vectors (x) –A has special structure (e.g., symmetric, banded, …) –Hybrid data structures (e.g., splitting, switch-to-dense, …) –Matrix reordering How and when do we search? –Off-line: Benchmark implementations –Run-time: Estimate matrix properties, evaluate performance models based on benchmark data

179. Cache Blocked SpMV on LSI Matrix: Ultra 2i A 10k x 255k 3.7M non-zeros Baseline: 16 Mflop/s Best block size & performance: 16k x 64k 28 Mflop/s

180. Cache Blocking on LSI Matrix: Pentium 4 A 10k x 255k 3.7M non-zeros Baseline: 44 Mflop/s Best block size & performance: 16k x 16k 210 Mflop/s

181. Cache Blocked SpMV on LSI Matrix: Itanium A 10k x 255k 3.7M non-zeros Baseline: 25 Mflop/s Best block size & performance: 16k x 32k 72 Mflop/s

182. Cache Blocked SpMV on LSI Matrix: Itanium 2 A 10k x 255k 3.7M non-zeros Baseline: 170 Mflop/s Best block size & performance: 16k x 65k 275 Mflop/s

183. Inter-Iteration Sparse Tiling (1/3) [Strout, et al., ‘01] Let A be 6x6 tridiagonal Consider y=A 2 x –t=Ax, y=At Nodes: vector elements Edges: matrix elements a ij y1y1 y2y2 y3y3 y4y4 y5y5 t1t1 t2t2 t3t3 t4t4 t5t5 x1x1 x2x2 x3x3 x4x4 x5x5

184. Inter-Iteration Sparse Tiling (2/3) [Strout, et al., ‘01] Let A be 6x6 tridiagonal Consider y=A 2 x –t=Ax, y=At Nodes: vector elements Edges: matrix elements a ij Orange = everything needed to compute y 1 –Reuse a 11, a 12 y1y1 y2y2 y3y3 y4y4 y5y5 t1t1 t2t2 t3t3 t4t4 t5t5 x1x1 x2x2 x3x3 x4x4 x5x5

185. Inter-Iteration Sparse Tiling (3/3) [Strout, et al., ‘01] Let A be 6x6 tridiagonal Consider y=A 2 x –t=Ax, y=At Nodes: vector elements Edges: matrix elements a ij Orange = everything needed to compute y 1 –Reuse a 11, a 12 Grey = y 2, y 3 –Reuse a 23, a 33, a 43 y1y1 y2y2 y3y3 y4y4 y5y5 t1t1 t2t2 t3t3 t4t4 t5t5 x1x1 x2x2 x3x3 x4x4 x5x5

186. Inter-Iteration Sparse Tiling: Issues Tile sizes (colored regions) grow with no. of iterations and increasing out-degree –G likely to have a few nodes with high out-degree (e.g., Yahoo) Mathematical tricks to limit tile size? –Judicious dropping of edges [Ng’01] y1y1 y2y2 y3y3 y4y4 y5y5 t1t1 t2t2 t3t3 t4t4 t5t5 x1x1 x2x2 x3x3 x4x4 x5x5

187. Summary and Questions Need to understand matrix structure and machine –BeBOP: suite of techniques to deal with different sparse structures and architectures Google matrix problem –Established techniques within an iteration –Ideas for inter-iteration optimizations –Mathematical structure of problem may help Questions –Structure of G? –What are the computational bottlenecks? –Enabling future computations? E.g., topic-sensitive PageRank  multiple vector version [Haveliwala ’02] –See www.cs.berkeley.edu/~richie/bebop/intel/google for more info, including more complete Itanium 2 results.

188. Exploiting Matrix Structure Symmetry (numerical or structural) –Reuse matrix entries –Can combine with register blocking, multiple vectors, … Matrix splitting –Split the matrix, e.g., into r x c and 1 x 1 –No fill overhead Large matrices with random structure –E.g., Latent Semantic Indexing (LSI) matrices –Technique: cache blocking Store matrix as 2 i x 2 j sparse submatrices Effective when x vector is large Currently, search to find fastest size

189. Symmetric SpMV Performance: Pentium 4

190. SpMV with Split Matrices: Ultra 2i

191. Cache Blocking on Random Matrices: Itanium Speedup on four banded random matrices.

192. Sparse Kernels and Optimizations Kernels –Sparse matrix-vector multiply (SpMV): y=A*x –Sparse triangular solve (SpTS): x=T -1 *b –y=AA T *x, y=A T A*x –Powers (y=A k *x), sparse triple-product (R*A*R T ), … Optimization techniques (implementation space) –Register blocking –Cache blocking –Multiple dense vectors (x) –A has special structure (e.g., symmetric, banded, …) –Hybrid data structures (e.g., splitting, switch-to-dense, …) –Matrix reordering How and when do we search? –Off-line: Benchmark implementations –Run-time: Estimate matrix properties, evaluate performance models based on benchmark data

193. Register Blocked SpMV: Pentium III

194. Register Blocked SpMV: Ultra 2i

195. Register Blocked SpMV: Power3

196. Register Blocked SpMV: Itanium

197. Possible Optimization Techniques Within an iteration, i.e., computing (G+uu T )*x once –Cache block G*x On linear programming matrices and matrices with random structure (e.g., LSI), 1.5—4x speedups Best block size is matrix and machine dependent –Reordering and/or splitting of G to separate dense structure (rows, columns, blocks) Between iterations, e.g., (G+uu T ) 2 x –(G+uu T ) 2 x = G 2 x + (Gu)u T x + u(u T G)x + u(u T u)u T x Compute Gu, u T G, u T u once for all iterations G 2 x: Inter-iteration tiling to read G only once

198. Multiple Vector Performance: Itanium

199. Sparse Kernels and Optimizations Kernels –Sparse matrix-vector multiply (SpMV): y=A*x –Sparse triangular solve (SpTS): x=T -1 *b –y=AA T *x, y=A T A*x –Powers (y=A k *x), sparse triple-product (R*A*R T ), … Optimization techniques (implementation space) –Register blocking –Cache blocking –Multiple dense vectors (x) –A has special structure (e.g., symmetric, banded, …) –Hybrid data structures (e.g., splitting, switch-to-dense, …) –Matrix reordering How and when do we search? –Off-line: Benchmark implementations –Run-time: Estimate matrix properties, evaluate performance models based on benchmark data

200. SpTS Performance: Itanium (See POHLL ’02 workshop paper, at ICS ’02.)

201. Sparse Kernels and Optimizations Kernels –Sparse matrix-vector multiply (SpMV): y=A*x –Sparse triangular solve (SpTS): x=T -1 *b –y=AA T *x, y=A T A*x –Powers (y=A k *x), sparse triple-product (R*A*R T ), … Optimization techniques (implementation space) –Register blocking –Cache blocking –Multiple dense vectors (x) –A has special structure (e.g., symmetric, banded, …) –Hybrid data structures (e.g., splitting, switch-to-dense, …) –Matrix reordering How and when do we search? –Off-line: Benchmark implementations –Run-time: Estimate matrix properties, evaluate performance models based on benchmark data

202. Optimizing AA T *x Kernel: y=AA T *x, where A is sparse, x & y dense –Arises in linear programming, computation of SVD –Conventional implementation: compute z=A T *x, y=A*z Elements of A can be reused: When a k represent blocks of columns, can apply register blocking.

203. Optimized AA T *x Performance: Pentium III

204. Current Directions Applying new optimizations –Other split data structures (variable block, diagonal, …) –Matrix reordering to create block structure –Structural symmetry New kernels (triple product RAR T, powers A k, …) Tuning parameter selection Building an automatically tuned sparse matrix library –Extending the Sparse BLAS –Leverage existing sparse compilers as code generation infrastructure –More thoughts on this topic tomorrow

205. Related Work Automatic performance tuning systems –PHiPAC [Bilmes, et al., ’97], ATLAS [Whaley & Dongarra ’98] –FFTW [Frigo & Johnson ’98], SPIRAL [Pueschel, et al., ’00], UHFFT [Mirkovic and Johnsson ’00] –MPI collective operations [Vadhiyar & Dongarra ’01] Code generation –FLAME [Gunnels & van de Geijn, ’01] –Sparse compilers: [Bik ’99], Bernoulli [Pingali, et al., ’97] –Generic programming: Blitz++ [Veldhuizen ’98], MTL [Siek & Lumsdaine ’98], GMCL [Czarnecki, et al. ’98], … Sparse performance modeling –[Temam & Jalby ’92], [White & Saddayappan ’97], [Navarro, et al., ’96], [Heras, et al., ’99], [Fraguela, et al., ’99], …

206. More Related Work Compiler analysis, models –CROPS [Carter, Ferrante, et al.]; Serial sparse tiling [Strout ’01] –TUNE [Chatterjee, et al.] –Iterative compilation [O’Boyle, et al., ’98] –Broadway compiler [Guyer & Lin, ’99] –[Brewer ’95], ADAPT [Voss ’00] Sparse BLAS interfaces –BLAST Forum (Chapter 3) –NIST Sparse BLAS [Remington & Pozo ’94]; SparseLib++ –SPARSKIT [Saad ’94] –Parallel Sparse BLAS [Fillipone, et al. ’96]

207. Context: Creating High-Performance Libraries Application performance dominated by a few computational kernels Today: Kernels hand-tuned by vendor or user Performance tuning challenges –Performance is a complicated function of kernel, architecture, compiler, and workload –Tedious and time-consuming Successful automated approaches –Dense linear algebra: ATLAS/PHiPAC –Signal processing: FFTW/SPIRAL/UHFFT

208. Cache Blocked SpMV on LSI Matrix: Itanium

209. Sustainable Memory Bandwidth

210. Multiple Vector Performance: Pentium 4

211. Multiple Vector Performance: Itanium

212. Multiple Vector Performance: Pentium 4

213. Optimized AA T *x Performance: Ultra 2i

214. Optimized AA T *x Performance: Pentium 4

215. Tuning Pays Off—PHiPAC

216. Tuning pays off – ATLAS Extends applicability of PHIPAC; Incorporated in Matlab (with rest of LAPACK)

217. Register Tile Sizes (Dense Matrix Multiply) 333 MHz Sun Ultra 2i 2-D slice of 3-D space; implementations color- coded by performance in Mflop/s 16 registers, but 2-by-3 tile size fastest

218. High Precision GEMV (XBLAS)

219. High Precision Algorithms (XBLAS) Double-double ( High precision word represented as pair of doubles) –Many variations on these algorithms; we currently use Bailey’s Exploiting Extra-wide Registers –Suppose s(1), …, s(n) have f-bit fractions, SUM has F>f bit fraction –Consider following algorithm for S =  i=1,n s(i) Sort so that |s(1)|  |s(2)|  …  |s(n)| SUM = 0, for i = 1 to n SUM = SUM + s(i), end for, sum = SUM –Theorem (D., Hida) Suppose F<2f (less than double precision) If n  2 F-f + 1, then error  1.5 ulps If n = 2 F-f + 2, then error  2 2f-F ulps (can be  1) If n  2 F-f + 3, then error can be arbitrary (S  0 but sum = 0 ) –Examples s(i) double (f=53), SUM double extended (F=64) – accurate if n  2 11 + 1 = 2049 Dot product of single precision x(i) and y(i) –s(i) = x(i)*y(i) (f=2*24=48), SUM double extended (F=64)  –accurate if n  2 16 + 1 = 65537

220. More Extra Slides

221. Opteron (1.4GHz, 2.8GFlop peak) Beat ATLAS DGEMV’s 365 Mflops Nonsymmetric peak = 504 MFlopsSymmetric peak = 612 MFlops

222. Awards Best Paper, Intern. Conf. Parallel Processing, 2004 –“Performance models for evaluation and automatic performance tuning of symmetric sparse matrix-vector multiply” Best Student Paper, Intern. Conf. Supercomputing, Workshop on Performance Optimization via High-Level Languages and Libraries, 2003 –Best Student Presentation too, to Richard Vuduc –“Automatic performance tuning and analysis of sparse triangular solve” Finalist, Best Student Paper, Supercomputing 2002 –To Richard Vuduc –“Performance Optimization and Bounds for Sparse Matrix-vector Multiply” Best Presentation Prize, MICRO-33: 3 rd ACM Workshop on Feedback-Directed Dynamic Optimization, 2000 –To Richard Vuduc –“Statistical Modeling of Feedback Data in an Automatic Tuning System”

223. Accuracy of the Tuning Heuristics (4/4)

224. Can Match DGEMV Performance DGEMV

225. Fraction of Upper Bound Across Platforms

226. Achieved Performance and Machine Balance Machine balance [Callahan ’88; McCalpin ’95] –Balance = Peak Flop Rate / Bandwidth (flops / double) –Lower is better (I.e. can hope to get higher fraction of peak flop rate) Ideal balance for mat-vec:  2 flops / double –For SpMV, even less

228. Where Does the Time Go? Most time assigned to memory Caches “disappear” when line sizes are equal –Strictly increasing line sizes

229. Execution Time Breakdown: Matrix 40

230. Execution Time Breakdown (PAPI): Matrix 40

231. Speedups with Increasing Line Size

232. Summary: Performance Upper Bounds What is the best we can do for SpMV? –Limits to low-level tuning of blocked implementations –Refinements? What machines are good for SpMV? –Partial answer: balance characterization Architectural consequences? –Help to have strictly increasing line sizes

233. Evaluating algorithms and machines for SpMV Metrics –Speedups –Mflop/s (“fair” flops) –Fraction of peak Questions –Speedups are good, but what is “the best?” Independent of instruction scheduling, selection Can SpMV be further improved or not? –What machines are “good” for SpMV?